Base-catalyzed oxyalkylation has been used to prepare polyoxyalkylene polyols for many years. In such a process, a suitably hydric low molecular weight starter molecule, such as propylene glycol or glycerine, is oxyalkylated with one or more alkylene oxides; such as ethylene oxide or propylene oxide, to form a polyoxyalkylene polyether polyol product. Because it is possible to employ a low molecular weight starter, the build ratio (polyol weight/starter weight) is relatively high, and thus the process effectively utilizes reactor capacity. Strongly basic catalysts such as sodium hydroxide or potassium hydroxide are typically used in such oxyalkylations.
Thus, most of polyoxyalkylene polyols useful in synthesis of polyurethane polymers, as well as those suitable for other uses, contain substantial amounts of oxypropylene moieties. As those skilled in the art are aware, during base-catalyzed oxypropylation, a competing rearrangement of propylene oxide to allyl alcohol generates monofunctional species which also become oxyalkylated, producing a wide range of polyoxyalkylene monols with molecular weights ranging from that of allyl alcohol itself or its low molecular weight oxyalkylated oligomers to polyether monols of very high molecular weight. In addition to broadening the molecular weight distribution of the product, the continuous generation of monols lowers the product functionality. For example, a polyoxypropylene diol or triol of 2,000 Da equivalent weight may contain from 30 to 40 mole percent monol. The monol content lowers the functionality of the polyoxypropylene diols produced from their “nominal” or “theoretical” functionality of 2.0 to “actual” functionalities in the range of 1.6 to 1.7. In the case of triols, the functionality may range from 2.2 to 2.4. As the oxypropylation proceeds further, the functionality continues to decrease, and the molecular weight growth rate slows. For these reasons, the upper practical limit for base-catalyzed polyoxypropylene polyol equivalent weight is just above 2,000 Da. Even at those modest equivalent weights, the products are characterized by low actual functionality and broad molecular weight distribution.
The monol content of polyoxyalkylene polyols is generally determined by measuring the unsaturation, for example by ASTM D-2849-69, “Testing of Urethane Foam Polyol Raw Materials”, as each monol molecule contains allylic termination. Levels of unsaturation of about 0.060 meq/g to in excess of 0.10 meq/g for based-catalyzed polyols such as those described above are generally obtained. Numerous attempts have been made to lower unsaturation, and hence monol content, but few have been successful.
In the early 1960's, double metal cyanide (“DMC”) complexes, such as the non-stoichiometric glyme complexes of zinc hexacyanocobaltate, were found which were able to prepare polyoxypropylene polyols with low monol contents, as reflected by unsaturation in the range of 0.018 to 0.020 meq/g. This represented a considerable improvement over the monol content obtainable by base catalysis.
In the 1970's, General Tire & Rubber Company, in U.S. Pat. No. 3,829,505, described the preparation of high molecular weight diols, triols etc., using double metal cyanide catalysts. However, the catalyst activity, coupled with catalyst cost and the difficulty of removing catalyst residues from the polyol product, prevented commercialization of the products.
In the 1980's, interest in such catalysts resurfaced, and improved catalysts with higher activity coupled with improved methods of catalyst removal allowed commercialization for a short time. The polyols also exhibited somewhat lower monol content, as reflected by unsaturations in the range of 0.015 to 0.018 meq/g. However, the economics of the process were marginal, and in many cases, improvements expected in polymer products due to higher functionality and higher polyol molecular weight did not materialize.
In the 1990's, DMC catalysts were developed with far greater activity than was theretofore possible. Those catalysts, described for example in U.S. Pat. Nos. 5,470,813 and 5,482,908, allowed commercialization of DMC-catalyzed polyether polyols by ARCO Chemical Company under the ACCLAIM tradename. Unlike the low unsaturation (0.015-0.018 meq/g) polyols prepared by prior DMC catalysts, these ultra-low unsaturation polyols often demonstrated dramatic improvements in polymer properties, although formulations were often different from the formulations useful with conventional polyols. These polyols typically have unsaturation in the range of 0.002 to 0.008 meq/g.
As those skilled in the art appreciate, one drawback of DMC-catalyzed oxyalkylation is the difficulty of using low molecular weight starters in polyether synthesis. Polyoxyalkylation of low molecular weight starters is generally sluggish, and often accompanied by catalyst deactivation. Thus, rather than employing low molecular weight starter molecules directly, oligomeric starters are prepared in a separate process by base-catalyzed oxypropylation of a low molecular weight starter to equivalent weights in the range of 200 Da to 700 Da or higher. Further oxyalkylation to the target molecular weight takes place in the presence of DMC catalysts. However, it is known to those skilled in the art that strong bases deactivate DMC catalysts. Thus, the basic catalyst used in oligomeric starter preparation must be removed by methods such as neutralization, adsorption, ion exchange, and the like. Several such methods require prolonged filtration of viscous polyol. The additional steps associated with catalyst removal from the oligomeric starter can add significant process time, and thus cost, to the overall process. Furthermore, the higher molecular weight of the starter lowers the build ratio of the process significantly, thereby decreasing reactor utilization.
Another drawback associated with oxyalkylation with DMC catalysts is that a very high molecular weight component is generally observed. The bulk of DMC-catalyzed polyol product molecules are contained in a relatively narrow molecular weight band, and thus DMC-catalyzed polyols exhibit very low polydispersities, generally 1.20 or less. However, it has been determined that a very small fraction of molecules, i.e., less than 1,000 ppm, have molecular weights in excess of 100,000 Da. This very small, but very high molecular weight, fraction is thought to be responsible for some of the anomalous properties observed with ultra-low unsaturation, high functionality polyols. These ultra high molecular weight molecules do not significantly alter the polydispersity, however, due to the extremely small amounts present.
U.S. Pat. Nos. 5,777,177 and 5,689,012, disclose that the high molecular weight “tail” in polyoxypropylene polyols may be minimized by continuous addition of starter (“CAOS”) during oxyalkylation. In batch and semi-batch processes, low molecular weight starter, e.g., propylene glycol or dipropylene glycol, is added continuously as the polyoxyalkylation proceeds rather than all being added at the onset. The continued presence of low molecular weight species has been found to lower the amount of high molecular weight tail produced, while also increasing the build ratio, because a large proportion of the final polyol product is derived from low molecular weight starter itself. Surprisingly, the polydispersity remains low, contrary to an expected large broadening of molecular weight distribution. In the continuous addition process, continuous addition of starter during continuous rather than batch production was found to also result in less low molecular weight tail, while allowing a build ratio which approaches that formerly obtainable only by traditional semi-batch processing employing base catalysis.
Unfortunately, when glycerine, a widely used trifunctional starter, is employed in either the batch-type continuous addition of starter process, or the continuous-type continuous addition of starter process, the DMC catalyst gradually deactivates, and often a polyether of the desired molecular weight cannot be obtained, or when obtained, product characteristics such as amount of high molecular weight tail, polydispersity, etc., are less than optimal. It appears that in the low molecular weight range of about 260 to 2500, where the ratio of glycerin to propylene oxide is higher than it is when making high molecular weight polyols, glycerin and other low molecular weight starters can act as inhibitors and stress the catalyst. Any other effects may be more evident under these stressed conditions. Because glycerine is derived from plant or animal matter by base-dependent processes, it contains one or more basic contaminants which may cause a loss of DMC catalyst activity. McDaniel et al., recognize this and teach in U.S. Pat. No. 6,077,978, the addition of very small amounts (i.e., up to about 100 ppm) of acid to the glycerine initiator prior to its introduction into the reactor as continuously added starter to neutralize the basic contaminants. Synthetic glycerin may have trace residues of base from the manufacturing process. Methods said to be useful other than the addition of acid, according to the '978 patent, include adsorption by acid adsorbents, and ion-exchange to either neutralize the impurities or to exchange them for acidic moieties. The addition of acid, however, is the preferred method of McDaniel et al. for increasing the DMC catalyst's ability to resist deactivation during CAOS feeds at high CAOS/oxide ratios. The '978 patent states at col. 6, lines 55-58, that, “ . . . less than 100 ppm acid based on total low molecular weight starter need be added, preferably about 5 ppm to 50 ppm, and most preferably about 10 ppm to 30 ppm” should be used. McDaniel et al. fail to provide any teaching or suggestion to add greater amounts of acid than that needed to neutralize basic contaminants of the glycerine.
It would be desirable to be able to utilize low molecular weight starter molecules for low molecular weight polyol production using DMC catalysis. It would further be desirable to prepare DMC-catalyzed polyols with minimal high molecular weight tail components. It would be further desirable to prepare polyoxyalkylation polyols in high build ratios. However, these objectives cannot be met if catalyst deactivation occurs.